Professor Wäfler, the term “sociotechnical system” is often used in connection with human-machine teaming. What does this mean exactly?
The term describes companies as systems consisting of technical and human parts. Together, they form a sociotechnical system. The technical subsystem includes the machines, the IT, but also processes, rules and regulations, and the physical infrastructure. The social subsystem includes the individual employees, the teams in which they work, and other subgroups with their individual needs and behaviors. This distinction is important because technical and social subsystems operate according to different laws. For a company to get the most out of this system, it has to coordinate the interplay between people, technology, and the organization, in the best possible way. If you only consider the technical side of the equation, the overall system is certain to function poorly.
So humans and machines are becoming colleagues. But how exactly should these human-machine combinations look in practice to avoid being “smart” in theory only?
When it comes to humans and artificial intelligence (AI) working together, it’s important that AI is capable of explaining itself in future – known as “explainable artificial intelligence.” This is the only way that humans will be able to work with it effectively. After all, even AI in the form of a robot or a decision-support system makes mistakes – and humans need to be able to identify them as such. But this is only possible if the AI explains itself, i.e. it provides arguments for its suggestions or actions.
So each side has its strengths. What can people do better than AI, and what can AI do better than people?
AI is good at calculations and can incorporate a lot of data. If it’s well trained, it can provide well-founded answers to the questions it’s specialized in. On the other hand, human beings can think and possess a wide range of knowledge, both specialized and general. This knowledge helps them interpret and validate AI’s proposals. This skill is extremely important. If, for example, the data that is used to train an AI is biased – which is not unlikely – then the AI’s responses will also contain errors. Humans can recognize that.
In other words – and in contrast to what we constantly hear in discussions about the jobs of the future – people don’t really need any new skills at all?
Well at the very least, specialized knowledge and a good general education will continue to be very important in the future. That’s why it’s sometimes misleading to talk about new skills. What is true is that demands will increase wherever technology is used to intelligently combine people and technology rather than replace people. As processes become increasingly complex, people need additional specialized knowledge and methodological skills. In addition, interdisciplinary work will become even more important than before. Take Formula 1 and Michael Schumacher, for example. They say he was especially good at understanding his car. That meant he was able to communicate with engineers effectively, and they were able to use the information to perfectly fine-tune the car. Today, it’s no longer enough to be a good driver – meaning being good in “your” field. You need to be able to communicate and work well with other specialists. Moreover, in contrast to AI, humans can apply their contextual knowledge and broad experience.
Many companies complain that it’s precisely this knowledge and practical experience that’s increasingly being lost. Could human-machine teaming be a possible solution?
Work has to be structured so that people can continue to gain practical experience. Throughout the history of work, people have acquired much of their knowledge simply by doing. If a machine takes over the doing entirely, humans can no longer gain experience. At the same time, complexity continues to increase. Knowledge gained from experience will only be preserved if future generations also gain relevant practical experience. No matter how much we end up automating, this aspect always needs to be taken into account. And in the end, companies benefit because experience and contextual knowledge are exactly what make humans so valuable because machines can’t replicate that.
You once said that people can only realize their full potential in human-machine teaming if Industry 4.0 technologies are introduced intelligently. What does that mean exactly?
We can’t view people and technology as competitors. Trying to use technology to copy people or even replace them is the wrong approach. It is possible, of course, but it’s not a smart combination. A smart combination is when we view people and technology as complementary. For example, in aviation, the interaction between pilot and aircraft is what makes it possible to travel from A to B. Even in highly automated aircraft equipped with state-of-the-art technology, this is still the case. Monitoring automated flight is a challenging task, as is intervening in the event of an emergency. Flying requires airports, air traffic controllers, aircraft maintenance – all complex sociotechnical systems. These systems can be transposed onto the entire world of work, like process manufacturing, for example. Even the smartest factory will still need people.
And what if I just create a system without people? Then I won’t have an interplay-problem, and machines also cost less money than people in the long run...
Despite all the progress and technological developments, I still simply can’t imagine such a scenario. Theories that lean in this direction are, in my opinion, wrong. There are certainly individual tasks that can be completely automated. But there will always be some tasks in which people are involved, like maintaining and advancing systems, for example. To do these tasks, people will need to have operational knowledge of how to run the system. If I were to create a system so complex that no one could understand it, no one would be able to monitor it and control it in an emergency. This means that overall, it couldn’t be operated with the necessary level of safety. So you can certainly take automation too far. Believing that you’re being economically efficient by completely replacing humans with machines is simply an illusion. On the contrary, sometimes it’s actually more economical not to automate to the absolute limit, but instead to find the right mix of human and machine.
What can companies do to ensure that people and technology work together in the best possible way?
We have to further refine the methods we already know – after all, human-machine interaction is nothing new – against the backdrop of new technological developments, and make them suitable for use in the real world. A company’s attitude is also important. Do I view my employees as a central asset, in the sense that they are a valuable source of knowledge and experience? Or do I simply see them as production factors that need to be automated away as much as possible? My position is clear: competitive advantages in this part of the world aren’t found in technology, but in the minds and hands of our employees. The competition can always acquire technology. The trick is using the technology better than the competition. This transforms employees into the key competitive factor, because you can’t simply copy the intelligent combination of people, technology, and organization. Preserving and expanding knowledge and experience are central to this. The use of technology can’t lead to people losing knowledge and skills. If one day an aircraft is completely automated, it can’t be at the expense of the pilot’s flying skills. To prevent this, companies need to understand exactly what the core competencies are in the minds and hands of their employees, and how to automate so that those skills aren’t lost.
You’re an expert in industrial psychology. From a psychological perspective, what effect does automation – and the resulting change in work – have on people?
Besides what I’ve mentioned so far, change makes people nervous. This is especially true when the change directly affects the employees themselves and they have to worry about losing control over what happens to them. This is when fears arise. We need to be careful that these fears don’t cause people to resist change. After all, companies are highly dependent on technological progress in order to remain competitive in the global marketplace. In other words, we must give as much control as possible to those affected by change by getting them involved in shaping the change.
That sounds all well and good – in theory. But how can companies actually implement a change process like this in the real world?
First, companies and their managers must understand that often it isn’t the new technology or automation that triggers resistance, or at least skepticism, but loss of control. Interestingly enough, excellent change management is still often regarded as nice-to-have, even though it’s an absolute must-have. Companies need to get their employees involved if the change is going to succeed. After all, it is precisely those employees who know the systems, machines, and processes. With their knowledge, they are by far the best and most effective way to manage change. If you don’t get them on board, they increasingly keep that knowledge to themselves because it protects them from being displaced by change. So you have to show that you’re committed to them – and you do that by involving them in the automation process. This doesn’t have anything to do with being nice, nor with the equally important ethical responsibility companies have for their employees. You can implement all the automation in the world, but new systems still need people to function.